Epilepsy Research (2014) 108, 1839—1844
journal homepage: www.elsevier.com/locate/epilepsyres
The relationship between hippocampal volumes and nonverbal memory in patients with medial temporal lobe epilepsy Bingwei Peng b,∗, Liwen Wu a, Lihua Zhang a, Yan Chen a a b
Department of Neurology, Peking Union Medical College Hospital, China Department of Neurology, Guang Zhou Women and Children’s Medical Center, China1
Received 15 May 2014; received in revised form 3 September 2014; accepted 6 September 2014 Available online 5 October 2014
KEYWORDS Temporal lobe epilepsy; Nonverbal memory; MRI volumetric analysis; Nonsense graphical recognition
∗ 1
Summary Objective: To explore the involvement of medial temporal lobe structures such as the hippocampus, amygdala, and entorhinal cortex (EC) in memory consolidation by volumetric magnetic resonance imaging (MRI). Methods: Sixty-two consecutive patients with medial temporal lobe epilepsy (MMTLE) were assessed using the Clinical Memory Scale (CMS) and MRI to measure the volumes of the hippocampus, amygdala, and EC. Participants were grouped according to MRI findings into 3 groups: left MRI-positive (abnormal hippocampal formation on the left side; n = 17), right MRIpositive (abnormal hippocampal formation on the left side; n = 9), and MRI-negative (normal hippocampal formation; n = 36). One-way analysis of variance (ANOVA) was used to assess group differences for all volumetric data (Z scores or asymmetry indexes (AI)), memory scale scores, and clinical parameters. Post hoc analyses were done with Fisher’s least significant difference (LSD) tests. AI = 100 × (L-R)/(L + R). ‘‘L’’ and ‘‘R’’ refer to the left and right volumes of each structure, respectively. Results: The nonsense graphical recognition tests and the facial memory tests were significantly different between the three groups. Post hoc analyses showed that the right MRI-positive group performed significantly worse than the MRI-negative group on nonsense graphical recognition tests (P = 0.008) and the left MRI-positive group had significantly lower scores than the MRInegative group on facial memory tests (P = 0.023). Conclusions: Nonverbal memory was correlated with the status of the right hippocampus. © 2014 Elsevier B.V. All rights reserved.
Corresponding author at: Department of Neurology, Guang Zhou Women and Children’s Medical Center, China. Tel.: +86 13060845449. E-mail addresses: [email protected] (B. Peng), [email protected] (L. Wu), [email protected] (L. Zhang). Current address.
http://dx.doi.org/10.1016/j.eplepsyres.2014.09.007 0920-1211/© 2014 Elsevier B.V. All rights reserved.
1840
Introduction Damage to the hippocampus and related medial temporal lobe (MTL) structures leads to memory deficits in patients with medial temporal lobe epilepsy (MTLE) (Leritz et al., 2006). The human medial temporal lobe is composed of the hippocampus, the amygdala, and the parahippocampal region. The entorhinal cortex (EC) is an important structure in the parahippocampal region. Volumetric MRI has previously been used to assess damage in the hippocampus, amygdala, and parahippocampal region in patients with MTLE (Watson et al., 1992; Bernasconi et al., 2003, 2005). MRI can reveal disease progression, identify seizure foci, and assist the diagnosis of MTLE. Volumetric MRI has also been used to investigate the relationship between hippocampal volumes and clinical neuropsychological memory test data (Reminger et al., 2004). Patients with left hippocampal pathologies performed more poorly on verbal memory tests (Sawrie et al., 2001). However, relatively few studies have examined the relationship between memory impairment and volumes of the amygdala or the parahippocampal region by MRI (Lillywhite et al., 2007; Alessio et al., 2006). Recent studies suggest that the EC controls the bidirectional flow of information to and from the hippocampus, and plays an important role in the declarative memory deficits of MTLE patients (Schwarcz and Witter, 2002). The amygdala mediates memory consolidation through norepinephrine (NE) release (McGaugh, 2000). Patients with MTLE were grouped according to the type of memory impairment they experienced—–verbal and nonverbal memory. In the majority of studies and verbal memory impairments have been clearly identified (AnderssonRoswall et al., 2004; Banos et al., 2004; Baker et al., 2003). Memory studies on the left or right anterior temporal lobe indicated that MTLE in the left brain clearly impairs verbal memory (Lee et al., 2002). The current understanding of the role of the right temporal lobe in memory is somewhat limited due to using the Warrington Recognition Memory Test for Faces as the only nonverbal memory measure. (McDermid Vaz, 2004). Impaired recognition of facial stimuli has clearly been documented concerning to the right temporal lobe (Barr, 1997) and nonsense figure recognition indicate involvement of the right temporal lobe in memory deficits. The aim of this study was to investigate the roles of various MTL regions including the hippocampus, amygdala, and EC in verbal and nonverbal memory in patients with MTLE. Patients were grouped according to hippocampal pathology after high-resolution MRI. We quantified the volumes of MTL structures and investigated potential correlations between memory measures and MRI data to evaluate relationships between the memory and these MTL structures.
Methods Participants Sixty-two consecutive patients with MTLE at the Epilepsy Center at Peking Union Medical College Hospital between July 2006 and March 2007 were enrolled in this study. Each patient was diagnosed with MTLE after a review of clinical findings that included complex partial seizures of temporal
B. Peng et al. lobe origin through ictal video electroencephalography (VEEG) in conjunction with interictal EEG findings, neuroimaging findings, and developmental and clinical history (International League Against Epilepsy ILAE, 1989). Inclusion criteria included: (1) chronological age from 14 to 70 years; (2) no other neurological or mental disorders; (3) normal intelligence with a full-scale intelligence quotient >80 (provided by the Neuropsychological Laboratory); (4) no observable MRI evidence of neocortical lesions; (5) right handedness. Twenty healthy controls matched for gender and age were also enrolled in the study. The ethics committee at our hospital approved this study and all participants gave written informed consent. Other epilepsy variables were recorded, such as treatment (polytherapy or monotherapy), occupation, and seizure frequency that included complex partial seizure per month (CPS/m) and the total lifetime secondary general tonic clonic seizures (SGTC).
Memory assessment All patients underwent memory assessments according to the Clinical Memory Scale (CMS) during their first visit to our hospital. Researchers were blinded to patient groups. The CMS tests a variety of memory functions including: (1) Directed memory where 2 series of words were presented by the recorder, and each test was 24 words consisting of 12 target words (animals or fruits) and 12 sham or distractor words. The subjects were asked to immediately recall the targets during a 2-min period. (2) Verbal paired associates recall where 12 pairs of words with or without logical relationships were shown and tested 3 times in different sequences. When examiner spoke the first word of each pair, the subjects were asked to recall the second word for each test. (3) Familiar picture recall that is similar to the visual recall portion of the Wechsler memory scale WMS-III. (4) Nonsense graphical recognition where each participant was presented with 20 target nonsense graphics sequentially (e.g. blocked curve, blocked straight line, straight and curve, non-block curve, and non-block straight line). The recognition testing was done immediately following the presentation. The patient was told they would be asked to recognize the 20 target graphics among 40 graphics, where 20 were targets and 20 were shams. (5) Facial memory was tested by the patient being shown 6 faces along with a corresponding last name, occupation, and hobby. Afterward, patients were asked to recall the information corresponding to each face. Recalling the correct last name gave the subject twice as many points as the correct occupation or hobby. Directed memory and verbal paired associates recall estimated verbal memory, while familiar picture recall and nonsense graphical recognition estimated nonverbal memory. In China, the CMS was administered to 2161 educated, healthy adults to obtain normal values. All scores in our study were based on normative data of age-matched controls and thus, the effect of age was excluded.
MRI acquisition Images were obtained on a 3.0 T MR scanner (Signa, General Electric), including baseline MRI during their first visit to
Relationship between hippocampal volumes and nonverbal memory
1841
our facility. All patients underwent horizontal T2-weighted imaging using the following parameters: echo train length, 19; repetition time (TR), 4000 ms; echo time (TE), 102 ms; number of excitations (NEX), 2; field of view (FOV), 24 cm; slice thickness, 5 mm. T1-weighted sagittal images were also acquired using a TR of 500 ms and a TE of 15 ms. Coronal images were oriented perpendicular to the long axis of the hippocampus and included two T1 fluid-attenuated inversion recovery (FLAIR) sequences with the following parameters: (1) TR, 2000 ms; TE, Min Full; inversion time (TI), 860 ms; NEX, 2; FOV, 24 cm; slice thickness, 3 mm; and (2) TR, 8000 ms; TE, 124 ms; TI, 2250 ms; NEX, 2; FOV, 24 cm; slice thickness, 3 mm. Three-dimensional spoiled-gradient (SPGR) echo images were also acquired with the following parameters: TI, 450 ms; flip angle, 15◦ ; NEX, 2; FOV, 24 cm; and slice thickness, 1.6 mm. All results were analyzed by a neurologist and a radiologist, if there were any disagreements in MRI interpretation, researchers convened to discuss and decide the final result. All patient names and dates of MRI images were hidden during analysis. MTLE patients were grouped according to hippocampal findings, including abnormal formation or signal, or atrophy. Participants with MTLE were designated as left MRI-positive (n = 17), right MRI-positive (n = 9), or MRI-negative (n = 36).
Volumetric analysis Volumetric analysis was performed using an interactive software package developed in the Neuroimaging Laboratory at the Montreal Neurological Institute. All researchers were blinded to patient identities and experimental groups. The hippocampus, amygdala, and EC were manually segmented from each 3-dimensional coronal image according to previously described protocols (Watson et al., 1992; Bonilha et al., 2005; Lambert et al., 2003) to obtain the area outlined (indicating the relevant structure) by the slice thickness. The total volume of the structure was then calculated by summing all slice volumes.
Anatomic guidelines The anatomic guidelines used to outline the amygdala, hippocampus, and EC were obtained from previous protocols (Watson et al., 1992; Bonilha et al., 2005; Lambert et al., 2003) (Fig. 1). The superior rim of the amygdala lies in the fundus of the semianular sulcus. Its medial border is covered by part of the entorhinal cortex. The superior border of the amygdala resides where the closure of the lateral sulcus forms the endorhinal sulcus. The inferior and lateral borders of the amygdala are formed by the inferior horn of the lateral ventricle or white matter. Its posterior end occupies the roof of the lateral ventricular horn. The hippocampus is separated from the amygdala by a line connecting the inferior horn of the lateral ventricle to the sulcus at the inferior margin of the semilunar gyrus. The anterior limit of the hippocampus is at a line where the lateral geniculate body and the fimbria become visible. The medial border of the hippocampus is the cisterna ambiens and its lateral border is the temporal horn of lateral ventricle. The tail of the hippocampus is located at the isthmus of the cingulated gyrus. The medial border of EC is the ventral border of the gyrus
Figure 1 Anatomic guidelines of the medial temporal structure. A) shows the amygdala (L) and the anterior portion of the entorhinal cortex (R) B) shows the amygdala (L), hippocampal head, and middle portion of the entorhinal cortex (R) C) shows the tip of the entorhinal cortex (L) and hippocampal body, including the hippocampal tail (R).
semilunaris. The uncal cleft is used as the border that separates the hippocampal head from the EC. The most anterior coronal slice where the body of the hippocampus became clearly visible was chosen as the posterior limit of the EC.
Data analyses and statistics Volume measurements from each patient were standardized to the value of normal controls using a Z-score transformation. Statistical tests were performed with SPSS software (version 11.0). The data for educational level, chronological age, and duration were found to have non-normal distributions following one-sample Kollomogorov—Smirnoff tests. Non-normal data were transformed to normal distributions using a log X transformation. Gender, treatment, and occupation were analyzed by 2 tests and differences in seizures were analyzed by a nonparametric test. One-way analysis of variance (ANOVA) was used to assess group differences for all
1842
B. Peng et al.
volumetric measurements, memory scale scores, and clinical parameters. Post hoc analyses were done with Fisher’s least significant difference (LSD) tests. P values
journal homepage: www.elsevier.com/locate/epilepsyres
The relationship between hippocampal volumes and nonverbal memory in patients with medial temporal lobe epilepsy Bingwei Peng b,∗, Liwen Wu a, Lihua Zhang a, Yan Chen a a b
Department of Neurology, Peking Union Medical College Hospital, China Department of Neurology, Guang Zhou Women and Children’s Medical Center, China1
Received 15 May 2014; received in revised form 3 September 2014; accepted 6 September 2014 Available online 5 October 2014
KEYWORDS Temporal lobe epilepsy; Nonverbal memory; MRI volumetric analysis; Nonsense graphical recognition
∗ 1
Summary Objective: To explore the involvement of medial temporal lobe structures such as the hippocampus, amygdala, and entorhinal cortex (EC) in memory consolidation by volumetric magnetic resonance imaging (MRI). Methods: Sixty-two consecutive patients with medial temporal lobe epilepsy (MMTLE) were assessed using the Clinical Memory Scale (CMS) and MRI to measure the volumes of the hippocampus, amygdala, and EC. Participants were grouped according to MRI findings into 3 groups: left MRI-positive (abnormal hippocampal formation on the left side; n = 17), right MRIpositive (abnormal hippocampal formation on the left side; n = 9), and MRI-negative (normal hippocampal formation; n = 36). One-way analysis of variance (ANOVA) was used to assess group differences for all volumetric data (Z scores or asymmetry indexes (AI)), memory scale scores, and clinical parameters. Post hoc analyses were done with Fisher’s least significant difference (LSD) tests. AI = 100 × (L-R)/(L + R). ‘‘L’’ and ‘‘R’’ refer to the left and right volumes of each structure, respectively. Results: The nonsense graphical recognition tests and the facial memory tests were significantly different between the three groups. Post hoc analyses showed that the right MRI-positive group performed significantly worse than the MRI-negative group on nonsense graphical recognition tests (P = 0.008) and the left MRI-positive group had significantly lower scores than the MRInegative group on facial memory tests (P = 0.023). Conclusions: Nonverbal memory was correlated with the status of the right hippocampus. © 2014 Elsevier B.V. All rights reserved.
Corresponding author at: Department of Neurology, Guang Zhou Women and Children’s Medical Center, China. Tel.: +86 13060845449. E-mail addresses: [email protected] (B. Peng), [email protected] (L. Wu), [email protected] (L. Zhang). Current address.
http://dx.doi.org/10.1016/j.eplepsyres.2014.09.007 0920-1211/© 2014 Elsevier B.V. All rights reserved.
1840
Introduction Damage to the hippocampus and related medial temporal lobe (MTL) structures leads to memory deficits in patients with medial temporal lobe epilepsy (MTLE) (Leritz et al., 2006). The human medial temporal lobe is composed of the hippocampus, the amygdala, and the parahippocampal region. The entorhinal cortex (EC) is an important structure in the parahippocampal region. Volumetric MRI has previously been used to assess damage in the hippocampus, amygdala, and parahippocampal region in patients with MTLE (Watson et al., 1992; Bernasconi et al., 2003, 2005). MRI can reveal disease progression, identify seizure foci, and assist the diagnosis of MTLE. Volumetric MRI has also been used to investigate the relationship between hippocampal volumes and clinical neuropsychological memory test data (Reminger et al., 2004). Patients with left hippocampal pathologies performed more poorly on verbal memory tests (Sawrie et al., 2001). However, relatively few studies have examined the relationship between memory impairment and volumes of the amygdala or the parahippocampal region by MRI (Lillywhite et al., 2007; Alessio et al., 2006). Recent studies suggest that the EC controls the bidirectional flow of information to and from the hippocampus, and plays an important role in the declarative memory deficits of MTLE patients (Schwarcz and Witter, 2002). The amygdala mediates memory consolidation through norepinephrine (NE) release (McGaugh, 2000). Patients with MTLE were grouped according to the type of memory impairment they experienced—–verbal and nonverbal memory. In the majority of studies and verbal memory impairments have been clearly identified (AnderssonRoswall et al., 2004; Banos et al., 2004; Baker et al., 2003). Memory studies on the left or right anterior temporal lobe indicated that MTLE in the left brain clearly impairs verbal memory (Lee et al., 2002). The current understanding of the role of the right temporal lobe in memory is somewhat limited due to using the Warrington Recognition Memory Test for Faces as the only nonverbal memory measure. (McDermid Vaz, 2004). Impaired recognition of facial stimuli has clearly been documented concerning to the right temporal lobe (Barr, 1997) and nonsense figure recognition indicate involvement of the right temporal lobe in memory deficits. The aim of this study was to investigate the roles of various MTL regions including the hippocampus, amygdala, and EC in verbal and nonverbal memory in patients with MTLE. Patients were grouped according to hippocampal pathology after high-resolution MRI. We quantified the volumes of MTL structures and investigated potential correlations between memory measures and MRI data to evaluate relationships between the memory and these MTL structures.
Methods Participants Sixty-two consecutive patients with MTLE at the Epilepsy Center at Peking Union Medical College Hospital between July 2006 and March 2007 were enrolled in this study. Each patient was diagnosed with MTLE after a review of clinical findings that included complex partial seizures of temporal
B. Peng et al. lobe origin through ictal video electroencephalography (VEEG) in conjunction with interictal EEG findings, neuroimaging findings, and developmental and clinical history (International League Against Epilepsy ILAE, 1989). Inclusion criteria included: (1) chronological age from 14 to 70 years; (2) no other neurological or mental disorders; (3) normal intelligence with a full-scale intelligence quotient >80 (provided by the Neuropsychological Laboratory); (4) no observable MRI evidence of neocortical lesions; (5) right handedness. Twenty healthy controls matched for gender and age were also enrolled in the study. The ethics committee at our hospital approved this study and all participants gave written informed consent. Other epilepsy variables were recorded, such as treatment (polytherapy or monotherapy), occupation, and seizure frequency that included complex partial seizure per month (CPS/m) and the total lifetime secondary general tonic clonic seizures (SGTC).
Memory assessment All patients underwent memory assessments according to the Clinical Memory Scale (CMS) during their first visit to our hospital. Researchers were blinded to patient groups. The CMS tests a variety of memory functions including: (1) Directed memory where 2 series of words were presented by the recorder, and each test was 24 words consisting of 12 target words (animals or fruits) and 12 sham or distractor words. The subjects were asked to immediately recall the targets during a 2-min period. (2) Verbal paired associates recall where 12 pairs of words with or without logical relationships were shown and tested 3 times in different sequences. When examiner spoke the first word of each pair, the subjects were asked to recall the second word for each test. (3) Familiar picture recall that is similar to the visual recall portion of the Wechsler memory scale WMS-III. (4) Nonsense graphical recognition where each participant was presented with 20 target nonsense graphics sequentially (e.g. blocked curve, blocked straight line, straight and curve, non-block curve, and non-block straight line). The recognition testing was done immediately following the presentation. The patient was told they would be asked to recognize the 20 target graphics among 40 graphics, where 20 were targets and 20 were shams. (5) Facial memory was tested by the patient being shown 6 faces along with a corresponding last name, occupation, and hobby. Afterward, patients were asked to recall the information corresponding to each face. Recalling the correct last name gave the subject twice as many points as the correct occupation or hobby. Directed memory and verbal paired associates recall estimated verbal memory, while familiar picture recall and nonsense graphical recognition estimated nonverbal memory. In China, the CMS was administered to 2161 educated, healthy adults to obtain normal values. All scores in our study were based on normative data of age-matched controls and thus, the effect of age was excluded.
MRI acquisition Images were obtained on a 3.0 T MR scanner (Signa, General Electric), including baseline MRI during their first visit to
Relationship between hippocampal volumes and nonverbal memory
1841
our facility. All patients underwent horizontal T2-weighted imaging using the following parameters: echo train length, 19; repetition time (TR), 4000 ms; echo time (TE), 102 ms; number of excitations (NEX), 2; field of view (FOV), 24 cm; slice thickness, 5 mm. T1-weighted sagittal images were also acquired using a TR of 500 ms and a TE of 15 ms. Coronal images were oriented perpendicular to the long axis of the hippocampus and included two T1 fluid-attenuated inversion recovery (FLAIR) sequences with the following parameters: (1) TR, 2000 ms; TE, Min Full; inversion time (TI), 860 ms; NEX, 2; FOV, 24 cm; slice thickness, 3 mm; and (2) TR, 8000 ms; TE, 124 ms; TI, 2250 ms; NEX, 2; FOV, 24 cm; slice thickness, 3 mm. Three-dimensional spoiled-gradient (SPGR) echo images were also acquired with the following parameters: TI, 450 ms; flip angle, 15◦ ; NEX, 2; FOV, 24 cm; and slice thickness, 1.6 mm. All results were analyzed by a neurologist and a radiologist, if there were any disagreements in MRI interpretation, researchers convened to discuss and decide the final result. All patient names and dates of MRI images were hidden during analysis. MTLE patients were grouped according to hippocampal findings, including abnormal formation or signal, or atrophy. Participants with MTLE were designated as left MRI-positive (n = 17), right MRI-positive (n = 9), or MRI-negative (n = 36).
Volumetric analysis Volumetric analysis was performed using an interactive software package developed in the Neuroimaging Laboratory at the Montreal Neurological Institute. All researchers were blinded to patient identities and experimental groups. The hippocampus, amygdala, and EC were manually segmented from each 3-dimensional coronal image according to previously described protocols (Watson et al., 1992; Bonilha et al., 2005; Lambert et al., 2003) to obtain the area outlined (indicating the relevant structure) by the slice thickness. The total volume of the structure was then calculated by summing all slice volumes.
Anatomic guidelines The anatomic guidelines used to outline the amygdala, hippocampus, and EC were obtained from previous protocols (Watson et al., 1992; Bonilha et al., 2005; Lambert et al., 2003) (Fig. 1). The superior rim of the amygdala lies in the fundus of the semianular sulcus. Its medial border is covered by part of the entorhinal cortex. The superior border of the amygdala resides where the closure of the lateral sulcus forms the endorhinal sulcus. The inferior and lateral borders of the amygdala are formed by the inferior horn of the lateral ventricle or white matter. Its posterior end occupies the roof of the lateral ventricular horn. The hippocampus is separated from the amygdala by a line connecting the inferior horn of the lateral ventricle to the sulcus at the inferior margin of the semilunar gyrus. The anterior limit of the hippocampus is at a line where the lateral geniculate body and the fimbria become visible. The medial border of the hippocampus is the cisterna ambiens and its lateral border is the temporal horn of lateral ventricle. The tail of the hippocampus is located at the isthmus of the cingulated gyrus. The medial border of EC is the ventral border of the gyrus
Figure 1 Anatomic guidelines of the medial temporal structure. A) shows the amygdala (L) and the anterior portion of the entorhinal cortex (R) B) shows the amygdala (L), hippocampal head, and middle portion of the entorhinal cortex (R) C) shows the tip of the entorhinal cortex (L) and hippocampal body, including the hippocampal tail (R).
semilunaris. The uncal cleft is used as the border that separates the hippocampal head from the EC. The most anterior coronal slice where the body of the hippocampus became clearly visible was chosen as the posterior limit of the EC.
Data analyses and statistics Volume measurements from each patient were standardized to the value of normal controls using a Z-score transformation. Statistical tests were performed with SPSS software (version 11.0). The data for educational level, chronological age, and duration were found to have non-normal distributions following one-sample Kollomogorov—Smirnoff tests. Non-normal data were transformed to normal distributions using a log X transformation. Gender, treatment, and occupation were analyzed by 2 tests and differences in seizures were analyzed by a nonparametric test. One-way analysis of variance (ANOVA) was used to assess group differences for all
1842
B. Peng et al.
volumetric measurements, memory scale scores, and clinical parameters. Post hoc analyses were done with Fisher’s least significant difference (LSD) tests. P values
